The 1st Insurance Data Science conference took place at Cass Business School, London on 16 July 2018.

Insurance Data Science conference, 16 July

• Download full conference programme

[09:00-10:00] Keynote 1: Gareth Peters

• Modelling longevity risk for retirement and pension planning [slides]

[10:00-11:00] Session 1: Pricing / Claims modelling

• Sparsity with multi-type lasso penalties - Tom Reynkens [slides]
• Statistical analysis of weather-related property insurance claims - Christian Rohrbeck [slides]
• Claims frequency modelling using telematics car driving data - Mario Wüthrich [slides]

[11:00-11:30] Coffee

[11:30-12:30] Session 2: Lightning talks

• Machine learning for actuaries: An introduction - NN, Steven Perkins [slides]
• Truncated regression models for the analysis of operational losses due to fraud: A high performance computing implementation in R - Alberto Glionna [slides]
• Using Random Forest to estimate risk profiles and probability of breakdowns - Lara A. Neira Gonzalez [slides]
• Simulating economic variables using graphical models - Aniketh Pittea [slides]
• RShiny at Qatar Re: A business case study - Marc Rierola [slides]
• PnC reinsurance modeling using Python and TensorFlow - Pauli Rämö [slides]
• ‘KSgeneral’ : A package for fast, exact, Komogorov-Smirnov goodness of fit testing - Senren Tan [slides]

[12:30-13:30] Lunch

[13:30-14:30] Session 3: Capital and exposure modelling

• Statistical learning for portfolio tail risk measurement - Michael Ludkovski [slides]
• Reverse sensitivity testing: What does it take to break the model? - Silvana M. Pesenti [slides]
• Using R for catastrophe modelling of cyber risks in (re)insurance - Benjamin C. Dean [slides]

[14:30-15:00] Session 4: Panel discussion

[15:00-15:30] Coffee

[15:30-16:30] Session 5: Business Case Studies

• SFCR automated analysis using scraping, text mining and machine learning methods for benchmarking and capital modelling - Aurelien Couloumy [slides]
• Machine learning and fairness for commercial insurance - Oliver Laslett [slides]
• Getting value out of machine learning - Javier Rodriguez Zaurin [slides]

[16:30-17:30] Keynote 2: Eric Novik

• Making decisions under uncertainty using Bayesian inference and Stan [slides]

[17:30-17:35] Annoucment 2019

• Announcement Insurance Data Science conference 2019 - Mario Wüthrich

[18:00-22:00] Reception drinks & Dinner

• Ironmongers’ Hall

Stan in Insurance Workshop, 17 July

[8:30-9:00] Registration

[9:00 - 10:30] Eric Novik

• Intro to Stan, including:
  • Coding linear regression to assess wine quality
  • Demonstrating important parts of the Stan program
  • Doing some basic posterior predicting checking
  • Introduction to calibration and model comparison
  • Introduction to making decisions with Bayesian models

[10:30 - 11:00] Coffee

[11:00 - 12:30] Paul-Christian Bürkner

• From classical GLMs to multi-level model
  • Comparing classical GLMs with bayesian GLMs using rstanarm
  • Building more complex multi-level models using brms
  • Examples from pricing and claims reserving

[12:30 - 13:30] Lunch

[13:30 - 14:30] Mick Cooney & Jake Morris

• Case studies from the insurance industry
  • Loss development curves in Stan (Mick Cooney)
  • Hierarchical compartmental reserving models (Jake Morris)

[14:30 - 15:00] Coffee

[15:00 - 17:00] Working in groups with support of the presenters

• Work on your own problems or work through on of the following examples:
  • Loss curve case study (Mick Cooney)
  • Extreme value case study (Aki Vehtari)
  • Various insurance related examples from Markus Gesmann
  • Examples from the brms vignettes (Paul-Christian Bürkner)

Sponsors

Gold Sponsors

RStudio

Verisk

Mango Solutions

Silver Sponsors

Mirai

Oasis Loss Modelling Framework

Milliman

Agrippa Data Consulting